PHIL 105 Lecture 18: Lecture 18
Document Summary
Questions: suppose a is correlated with b in population p, does it follow that b is correlated with. A in population p? answer: no this doesn"t follow: if a is correlated with b, does it follow that a causes b? answer: no this doesn"t follow. Establishing a correlation is not suf cient for establishing causation. This should be apparent from the fact that if a is correlated with b, then b is correlated with a. Correlation statements amount to conjunctions of two simple statistical statements, so two surgery arguments can be used to support them. This is indirectly supported by the background information): target correlation in sample (2,3) (the percentage of a that are b and the percentage of a that are not b). !1: representativeness premise (indirectly supported by the background information), final conclusion (4,5) (proving that there is a correlation between for example, a and b). The same considerations apply here as to simple surgery arguments.